on a fast-track: human discovery, exploration and ... · to upper paleolithic transition in western...
TRANSCRIPT
17Number 74, June 2012
Forum
colonisation of the Bismarck Archipelago: New evidence from New Britain.
Archaeology in Oceania 39:101-130.
Turney, C.S.M., M. Bird, L.K. Fifield, R.G. Roberts, M. Smith, C.E. Dortch, R. Grun,
E. Lawson, L.K. Ayliffe, G.H. Miller, J. Dortch and R.G. Cresswell 2001b Early
human occupation at Devil’s Lair, southwestern Australia, 50,000 years ago.
Quaternary Research 55:3-13.
Turney, C.S.M., M.I. Bird and R.G. Roberts 2001a Elemental δ13C at Allen’s Cave,
Nullabor Plain: Assessing post-depositional disturbance and reconstructing
past environments. Journal of Quaternary Science 16:779-784.
Turney, C.S.M., A.P. Kershaw, S.C. Clemens, R. Branch, P.T. Moss and L.K. Fifield
2004 Millennial and orbital variations of El Niño/Southern Oscillation and
high latitude climate in the last glacial period. Nature 428:306-310.
Ugan, A. 2005 Does size matter? Body size, mass collecting and their implications
for prehistoric foraging behavior. American Antiquity 70:75-90.
Ugan, A., J. Bright and A. Rogers 2003 When is technology worth the trouble?
Journal of Archaeological Science 30:1315-1329.
Ulm, S. 2011 Coastal foragers on southern shores: Marine resource use in northeast
Australia since the late Pleistocene. In N.F. Bicho, J.A. Haws and L.G. Davis
(eds), Trekking the Shore: Changing Coastlines and the Antiquity of Coastal
Settlement, pp.441-461. Interdisciplinary Contributions to Archaeology. New
York: Springer Science+Business Media.
van Holst Pellekaan, S. 2011 Genetic evidence for the colonisation of Australia.
Quaternary International doi:10.1016/j.quaint.2011.04.014.
Veth, P., M. Smith, J. Bowler, K. Fitzsimmons, A. Williams and P. Hiscock 2009
Excavations at Parnkupirti, Lake Gregory, Great Sandy Desert: OSL ages for
occupation before the Last Glacial Maximum. Australian Archaeology 69:1-10.
Walshe, K. 1998 Taphonomy of Mungo B assemblage: Indicators for subsistence
and occupation of Lake Mungo. Archaeology in Oceania 33:200-206.
Wang, X., S. van der Kaars, A.P. Kershaw, M. Bird and F. Jansen 1999 A record
of fire, vegetation and climate through the last three glacial cycles from
Lombok Ridge Core G6-4, eastern Indian Ocean, Indonesia. Palaeogeography,
Palaeoclimatology, Palaeoecology 147:241-256.
Wang, Y.J., H. Cheng, R.L. Edwards, Z.S. An, J.Y. Wu, C-C. Chen and J.A. Dorale 2001
A high-resolution absolute-dated late Pleistocene monsoon record from Hulu
Cave, China. Science 294:2345-2348.
Warner, R.F. 1986 Hydrology. In D.N. Jeans (ed.), Australia: A Geography, Vol.1 (2nd
ed.), pp.49-79. Sydney: Sydney University Press.
Weninger, B. and O. Joris 2008 A 14C age calibration curve for the last 60 ka: The
Greenland-Hulu U/Th time scale and its impact on understanding the Middle
to Upper Paleolithic transition in western Eurasia. Journal of Human Evolution
55:772-781.
White, J.P. 1977 Crude, colourless and unenterprising: Prehistorians and their
views on the stone age of Sunda and Sahul. In J. Allen, J. Golson and R. Jones
(eds), Sunda and Sahul: Prehistoric Studies in Southeast Asia, Melanesia and
Australia, pp.13-30. London: Academic Press.
Wickler, S. and M. Spriggs 1988 Pleistocene human occupation of the Solomon
Islands. Antiquity 62:703-706.
Williams, M., E. Cook, S. van der Kaars, T. Barrows, J. Shulmeister and P. Kershaw
2009 Glacial and deglacial climatic patterns in Australia and surrounding
regions from 35,000 to 10,000 years ago reconstructed from terrestrial and
near-shore proxy data. Quaternary Science Reviews 28:2398-2419.
Zeanah, D.W. 2004 Sexual division of labor and central place foraging: A model for
the Carson Desert of western Nevada. Journal of Anthropological Archaeology
23:1-32.
ON A FAST-TRACK: HUMAN dISCOVERy, ExPLORATION ANd SETTLEMENT OF SAHUL
Jon M. ErlandsonDepartment of Anthropology and Museum of Natural and
Cultural History, University of Oregon, Eugene OR 97403, USA
Recent decades have seen the emergence of new scientific
paradigms for the origins of anatomically modern humans
(AMH), a series of relatively rapid migrations that took these
people from Africa to the far corners of the world, and the role
of coastlines, boats and fishing in facilitating such dispersals.
These paradigm shifts all meet in Sunda, Sahul and western
Melanesia, where archaeology and anthropological genetics
are illuminating some of the earliest evidence for seafaring and
maritime migrations by AMH. The exact timing, routes and
adaptive diversity involved in such migrations remain to be
fully documented.
O’Connell and Allen argue that the peopling of Sahul was
an ‘archaeologically instantaneous’ process occurring between
ca 46,000 and 44,000 ya. They offer a ‘speculative model’ that
deals parsimoniously with numerous complex issues. I have no
qualms about rapid coastal migrations, as coastlines offer linear
migration routes, entirely at sea-level and with few geographic
barriers, often with very similar marine resources over vast areas.
For the Pacific Rim, Erlandson et al. (2007) argued that maritime
skills and similarities in the ecology of kelp forest and other
coastal habitats may have facilitated a rapid coastal migration
and colonisation from Asia to the Americas shortly after the end
of the LGM. A similar model could apply to the margins of the
Indian Ocean, where mangroves, coral reefs and other nearshore
ecosystems support a diverse array of similar shellfish, fish, birds
and marine mammal taxa from Africa to Asia and Australia.
I find nothing in O’Connell and Allen’s ‘fast track’ migration
that is not eminently reasonable. I agree that such a colonisation
process almost certainly involved relatively sophisticated
watercraft and seafaring capabilities. I also agree that maritime
immigrants would have spread rapidly into resource-rich interior
areas. Coastal habitats are partly terrestrial, after all, and coastal
peoples worldwide are well-attuned to subsistence opportunities
in the adjacent interior. As O’Connell and Allen note, moreover,
not all coastlines are equally productive, so those offering fewer
aquatic or terrestrial resources might have been skipped over
to focus on coastal sweet spots described by Bulbeck (2007)
and others. O’Connell and Allen’s model may overestimate the
vulnerability of estuarine ecosystems to sea-level fluctuations,
as intertidal organisms and communities tend by nature to be
highly resilient. During a critical period (ca 50,000-35,000 BP)
for their model, moreover, Pope and Terrell (2008:8) suggested
that estuarine habitats were relatively extensive in South and
East Asia.
O’Connell and Allen may be correct that the colonisation of
Sahul occurred very rapidly, but understanding a late Pleistocene
human colonisation of Sahul – as well as other Pleistocene
coastal migrations – is clouded by significant problems
(Erlandson 2001, 2010). First, the colonisation process took
place at, near, or just beyond the effective range of 14C dating,
18 Number 74, June 2012
Forum
which reduces the confidence we can have in 14C dates from
archaeological sites, especially those run without the advanced
extraction, purification and measurement protocols essential for
dating samples of such age. Second, migrations from Sunda to
Sahul – primarily coastal and maritime in nature – took place
at a time when global sea-levels were ca 60-30 m below modern.
With sea-levels now at a 120,000 year high, vast areas of Sahul’s
coastal lowlands have been submerged by rising seas, including
most of the ancient coastlines early immigrants would have
traversed. Third, thousands of islands in Southeast Asia have still
seen relatively little research on Pleistocene archaeological sites:
a region that stunned the world with a relatively recent discovery
of Homo floresiensis may have more secrets to unveil. As a
result, I continue to teach my students that Sahul was probably
colonised ca 50,000±5000 years ago, an Aristotlean ‘golden mean’
that averages models by scholars who know much more than I
about the archaeology of the region.
There are some interesting parallels in O’Connell and Allen’s
model with a Clovis-First model for colonising the Americas
that is now in full retreat. Clovis-First advocates long argued for
a terminal Pleistocene terrestrial migration that settled North
and South America within roughly a millennium. They created
demographic models that were hypothetically possible, but now
are proven to have been wrong. Current archaeological and
genetic data suggest that the settlement of the Americas, which
may have involved a coastal migration and detours that saw
humans follow productive river systems deep into continental
interiors, took several millennia to accomplish. The American
experience, where two large continents now appear to have
been colonised, explored and settled within as little as three
millennia, could support O’Connell and Allen’s fast-track model
for colonising Sahul. One lesson learned from the widespread
distribution of Clovis sites in North America, however, is that
it took 2000 to 3000 years for pre-Clovis populations to reach
levels of archaeological visibility that also seem characteristic of
the earliest sites throughout Sahul.
Given some of the challenges alluded to above (drowned
shorelines, limited chronological resolution, lack of research
in key areas etc.), I hope the excessive conservatism of Clovis-
First models will not be repeated in Australia and Southeast Asia.
Given the active research and lively debate that continues on the
human colonisation of Sunda and Sahul, this seems unlikely. I
congratulate O’Connell and Allen on a fast track model for
the peopling of Sahul that is a reasonable hypothesis, one with
profound implications and numerous corollaries that provide
fertile ground for further research and testing.
SHOULd I STAy OR SHOULd I gO NOW? THE SPATIAL dyNAMICS OF FORAgINg ANd dIMINISHINg RETURNS
Brian F. Codding and Rebecca Bliege BirdDepartment of Anthropology, Stanford University, Stanford CA
94305, USA [email protected] [email protected]
O’Connell and Allen should be praised for their concise and
focused synthesis of the debates and issues surrounding the
colonisation of Sahul. As they have done for the timing of this
event (e.g. Allen and O’Connell 2003; O’Connell and Allen 2004),
they now reframe the scope and direction of research towards
understanding how and why people rapidly entered the continent
and where they went. We would like to use this opportunity to
expand upon the theoretical framework presented in this and
companion papers (Allen and O’Connell 2010; O’Connell et
al. 2010), by examining foraging decisions across three explicit
spatial scales and proposing some additional archaeological
predictions within each.
At the heart of their model is the assumption that the rate
at which foragers can capture food diminishes with foraging
time (Charnov 1974): the more per capita time spent exploiting
a particular resource, remaining within a particular patch or
continuing to forage within a single catchment (sometimes
termed habitat), the lower the expected gains will be (see Figure
1). According to O’Connell and Allen, this is the engine that
drives people into and across Sahul: when resources are depleted,
people are better off moving into adjacent, unoccupied areas
than staying within depleted environments. However, the three
different models from which these predictions derive (prey
choice, patch choice and ideal free distribution) all operate at
explicitly different scales represented as resources, patches and
catchments, respectively. How these respond to population
pressure at these different scales, and how people then respond
to those changes will determine how, where and whether people
are expected to move. The effects of foraging on the environment
build cumulatively across these spatial scales, and devising
testable predictions about mobility patterns depends on the scale
at which foraging optimality is analysed.
Within a patch, foragers should pursue the highest ranking
resource to the exclusion of all others until the point when
continuing to search for that resource is less profitable than
taking the next highest ranking resource on encounter. As
encounter rates with higher ranked resources decline through
foraging-related resource depression, this results in an ever
widening ‘diet breadth’, leading to ever increasing handling costs,
but likely reduced search costs (Figure 1a). The prey choice
model predicts that people should respond to reductions in
high ranked prey with a reduction in mobility and a focus on
technological enhancements to reduce the costs of handling
lower ranked resources in patch, the standard model of resource
intensification. As O’Connell and Allen point out, there is little
archaeological evidence for ‘broad-spectrum’ intensification
in Sahul until the late Holocene, suggesting that foragers did
not tolerate in-patch declines for long. If, as they suggest, high
ranked but slow to reproduce rocky shore resources were driving
the decline, we should not necessarily see changes in species
composition, but are likely to see reductions in the size of staple
shellfish species over time.
If foragers decide not to remain and intensify their foraging
effort, they may abandon their patch and move on to exploit
other patches. These decisions are crucial if resources are
distributed heterogeneously across different patches. In these
circumstances, the critical foraging decisions involve which
patch-type to exploit and how long to stay within a particular
patch. With increasing consumption pressure, high ranking
patches would decline in productivity, leading foragers to enter
lower ranked patches (see Figure 1b). At the patch scale, foragers
expand their range within a catchment or habitat to encompass
patches of lower quality resources as high quality patches decline.
19Number 74, June 2012
Forum
Expansions in foraging radii are likely accompanied by increases
in logistic mobility, with people moving further to access more
remote patches, identifiable by the deposition of exotic resources
or raw materials acquired from patches located at a distance
from the central place.
As with resources and patches, increased human consumption
pressure should decrease overall foraging efficiency within
catchments (Fretwell and Lucas 1969), eventually causing
people to resettle in new areas (Figure 1c). Being the highest
spatial scale, these movements are significantly greater than
movements within or across patches. Given the marginal costs of
movement into adjacent territories that remain unoccupied, this
would likely result in the rapid dispersal of people as discussed
by O’Connell and Allen. Foraging decisions at this scale form
the backbone of O’Connell and Allen’s argument. As a next step,
we encourage the development of a spatially explicit ideal free
distribution model (e.g. Winterhalder et al. 2010), which would
rank catchments based on the aggregate returns and diminishing
gains within geographically specific patches according to the
expected resources within those patches. While such a model
has not yet been developed for any continent in any context,
it would provide detailed predictions about where the earliest
settlements are likely to occur and where we might expect greater
populations throughout time (Birdsell 1953).
Depletions at different spatial scales also likely occur over
differing temporal scales. However, in circumstances where there
are few opportunity costs to movement – as in scenarios where
people have watercraft – individuals will likely relocate across
vast distances in order to occupy highly ranked catchments.
This may appear as pulses and waves of settlement, with the
first settlers inhabiting higher ranking catchments, exploiting
the highest ranking patches within those catchments, and
pursuing the highest ranking resources within those patches.
Determining how these patterns encourage movement requires
building models incorporating individual decision-making to
test predictions at the largest spatial scales. O’Connell and Allen
have made great progress with this and companion publications,
and based on the available data, they are certainly correct. Next
steps will include further refinement of the model particulars
and tests of its predictions with new archaeological data.
FOUR qUESTIONS ABOUT FORAgINg MOdELS ANd THE PROCESS OF COLONISATION
Iain DavidsonIDHA Partners, 10 Cluny Rd, Armidale NSW 2350, Australia
The target paper takes the debate about the narrative of
Australian archaeohistory a significant step forward, and sets up
some new research problems to be tackled.
O’Connell’s research with the Alyawara (Iliaura) (O’Connell
and Hawkes 1981) demonstrated that, despite their access to
purchased flour, modern fisher-gatherer-hunters will collect
seeds and grind them, provided there is an anthropologist who
can use a vehicle to drive them to the grasses. While this sounds
dismissive, it is actually very important for two reasons: (1) the
gatherers needed to know where and when the grasses were
suitable for harvest; (2) the anthropologist’s vehicle reduced the
cost of travel and search effectively to zero, altering the values
in the patch choice model. A distinguished ethnographer of
fisher-gatherer-hunters protested angrily about this work: ‘My
people do not forage optimally.’ I wondered, silently, whether
they were somehow more virtuous because they had not reached
optimality or perhaps they were somehow better than optimal.
This questioning also has implications: (3) are there behaviours
which do reduce the ‘optimality’ of foraging; and, (4) on what
time scales do the considerations of behavioural ecology have
to operate?
1. The Importance of InformationAll HBE/OFT models assume that the foragers have information
required to make the appropriate choices, and ethnography
demonstrates prodigious knowledge of plant and animal species
(e.g. O’Connell et al. 1983), as well as key indicators of seasonality.
Golson (1971) showed that there were some similarities between
the vegetation communities on either side of the Wallace Line,
but only small numbers of species familiar to initial colonists
of restricted distribution within Australia. New colonists needed
to acquire information about the nature of suitable resources,
especially as some contain toxins, whatever their means of
dealing with them (e.g. Beck 1992).
As a result of this uncertainty about information, O’Connell
and Allen are right in the way that Bowdler (1977) was right:
the easiest option was to focus on familiar resources that were
present on both coasts. As coastal populations expanded, so there
Resources
Patches
Catchments
- Population Impact +
- O
vera
ll Fo
ragi
ng R
etur
ns
+
12
3
1
2
3
(a)
(b)
(c)
Figure 1 Predicted foraging dynamics across each spatial scale: (a) in patch; (b) between patches; and, (c) across catchments. The arrows indicate the point at which foragers should switch to alternative resources, patches or catchments.
20 Number 74, June 2012
Forum
was pressure to move inland; water and its variable availability
was the chief limiting factor in many regions (cf. White and
O’Connell 1982:51).
The first colonists of any region needed information on plant
species, including variation in their availability and food yield.
New colonists needed to discover what the resources were, as
well as their relative value. Both availability and ranks of food
resources could change stochastically from season to season, year
to year and decade to decade, as well as biozone to biozone. In
the fullness of time, such information could be obtained, but,
as oral histories seem to suggest, passing on information about
variation was difficult when the time scales of variation were
longer than individual experience (Davidson 1992).
2. The Costs of ForagingThere are also hidden costs of foraging, such as the embedded
cost of O’Connell’s vehicle. These feature in the discussion
of risk reduction in Australian archaeohistory deriving from
the work of Oswalt (1976), who characterised technologies
in terms of the number of technounits. Oswalt’s fallacy is
exposed by his equating of a digging stick and a chimpanzee’s
termiting wand because they are both single sticks (McGrew
1987). The digging stick, unlike the termiting wand, requires
stone tools to make it and can be used many times. So, one of
the costs of human foraging is not only about food sources,
but also acquiring information about, and obtaining, stone for
tools. It is noteworthy that at Sandy Creek 1 in Cape York, the
earliest assemblages (ca 37 cal ka) were made on quartz crystal
(Morwood et al. 1995), as if the first people had not yet found the
raw material sources they later relied on.
3. Are there Adaptive Behaviours that Reduce the Optimality of Foraging?
Historically, symbolically supported sets of relationships and long
distance interactions were part of the mechanism by which such
uncertainties were dealt with, a process that began early (Veth et
al. 2011). This could have been a disadvantage, particularly by
defining ritual relationships with place that may have endured
beyond the time that such places were sweet spots. Given that
rock art of all ages, and much of the symbolic expression of
relationships in the historical period, involved relationships
with animals (and much less with plants), such expressions of
value might affect assessments of resource ranking. In addition,
early beads suggest symbolic marking of interpersonal relations
that could promote the adaptive success of populations that
differentiated in such ways (Boyd and Richerson 1987).
4. On what Time Scales does Behavioural Ecology Operate?
At some level, the sorts of decisions O’Connell and Allen address
are almost encapsulated by the choice of restaurant alluded to
in their title, but human behaviour was subject to many more
selective pressures than those individual choices. As people
exhausted resources or populations became too big for the
resources of the region, they could have moved on or targeted
lower ranked resources, perhaps becoming strategic generalists.
Symbolic attachment to places probably tethered some groups
who adjusted to the stress of variations in resources and gave up
the option of seeking higher ranked resources in the restaurant
over the next sand hill. Sometimes they failed, as the patterning
of radiocarbon dates suggests, but in the end they found ways to
cope with the scale of variation that the environment threw at
them (Davidson in press).
COMMENT
Tom D. DillehayVanderbilt University, 211 Kirkland Hall, Nashville TN 37240, USA
O’Connell and Allen have provided an articulate,
multidimensional model of a subject made complex by its
interdisciplinary nature, its long history of debate and the
continuing shortage of hard archeological evidence for many
regions. I would like to expand briefly on some of the more
general archaeological and anthropological aspects of the model,
which are necessarily dealt with rather cursorily in the review.
O’Connell and Allen discuss the implications of an
archaeological record defined primarily by infrequent, short-
term occupation sites produced by small, highly mobile groups.
They also infer that human populations were generally ‘small and
patchily distributed across Sahul’. Not discussed by the authors
are the implications of this record for the absence, or minimal
presence, of more complex hunter-gatherers at the end of the
Pleistocene period. In other continents, the relationship between
late Pleistocene plant communities and humans, for instance,
often laid the developmental foundations for long distance
social networks, reduced territoriality if not semi-sedentism, and
eventually plant domestication. In the northern Peruvian desert
and surrounding environs during the late Pleistocene, foragers
were experimenting with new food procurement strategies,
occasionally rejecting some or all of them, and then returning to
prior ones (Dillehay 2011). Although the archaeological record
of this region indicates short-term site occupation similar to
that described for much of Sahul, it also reflects continuous
use of the same habitats over several millennia, as evidenced by
repeated yet ephemeral occupation of the same site locales. Some
foragers even managed plant and animal resources by fostering
environmental conditions (e.g. perhaps intentional burning or
predator overkill) that promoted preferred foods. This certainly
was the case for palm nuts in the Amazon rainforest and for
shellfish along the Pacific coast of South America. Is there any
evidence of incipient foraging complexity or early resource
management in Sahul at the end of the Pleistocene, and, if so,
how might they have impacted patch choice, marginal value and ideal free distribution?
Although the O’Connell and Allen model is focused on
various concepts developed within optimal foraging strategy, I
wonder why the social aspects of late Pleistocene humans are not
discussed as possible factors influencing resource exploitation,
diet choice, site patterning and so forth. For parts of South
America, we can now infer that social fissioning was probably
less common than we had thought and that it occurred when the
local carrying capacity had been exceeded or social tensions were
high. When it did occur, which probably resulted in smaller social
units moving around the landscape, we suspect that splinter
groups maintained close ties with previous groups for economic
support and social interactions. We now understand that some
21Number 74, June 2012
Forum
early foraging societies institutionalised practices that served
additional (and diverse) functions for their individual members
and which were organised as relatively specific entities in
different places (e.g. Rowley-Conwy 2001). These practices were
ritual orders, social networks, technological traditions, perhaps
gender-based units, and others that were performed in specific
private or public places, such as individual huts or rock art sites.
Given the rich rock art of many areas of Sahul, can anything be
inferred from it regarding early social and behavioral patterns?
A point I find intriguing relates to diet breadth and the
complexity of technology. The review notes that a bifacial stone
tool technology (and other technologies?) may have been lost
en-route from Africa and South Asia to Sahul ‘due to a series of
demographic “bottlenecks” that ultimately reduced the cultural
transmission and thus the technological repertoire of early
colonisers.’ Rather than imposing cultural transmission as a
causal factor here, O’Connell and Allen suggest that diet breadth
and technological complexity may correlate to explain the lack
of diversity/complexity in late Pleistocene stone tool industries.
They argue that a wider diversity of exploited resources increases
‘handling time’ and procurement techniques, which leads to
greater technological complexity. In South America, the reverse
seems to be the case: simple and less complex unifacial industries
are generally correlated with broad-spectrum diets derived
from large and small game hunting and the gathering of a wide
variety of plant species. These diets required more handling and
allocation time and procurement strategies across many different
environments. The primary adaptive technology for these diets
was a unifacial industry, which provided tools used for a wide
variety of handling and other functions. More complex and
diverse technologies, on the other hand, are associated with
specialised game hunting. Further, the general pattern in Peru
was ongoing frequency shifts in the use of alternative adaptive
strategies that incorporated changes in the use of different
technologies and tool types (stemmed and fluted point and
unifacial industries) in different time periods and in different
ecological settings.
Along similar lines, it seems that bifacial technologies were
occasionally ‘lost’, or at least less diversified, as late Pleistocene
people spread through Central America and into the wide range
of environments in South America. But there are archaeological
records across the continent suggesting that these technologies
were both lost and recovered from time-to-time. We do not
know if the lost, narrowed or recovered technologies relate to
cultural transmission, but, if so, we need to examine early social
networks and how information was passed along them, and
whether ‘technological diversity is likely more a function of
ecological context and foragers’ intellectual solutions.’ This topic
requires more inter-continental comparative research.
COMMENT
David HortonIndependent Scholar [email protected]
There is much to like in this paper, such as putting together all
the most recent research and applying it to the age old problem
of trying to define the timing and nature of the first colonisation
of Australia. When Birdsell, and even when Bowdler and I,
considered the subject in necessarily broad terms, there was far
less data than there is now.
I am also delighted to see recognition of the importance
of climate change, and the dismissal of firestick farming and
megafaunal extinctions nonsense. It is long past time Australian
archaeology left these ancient shibboleths behind. They have
been a drag on the discipline, preventing the analysis of new
data in ideologically uncluttered ways. Worse, they have led to
the great increase in prescribed burning, and to an inability to
use past climate change as an analogue, and a warning, for what
the country faces as the planet warms.
But, and there must be one, I do have a general problem with
the paper. Many years ago Mortimer Wheeler famously said ‘as
for archaeology, I do not know whether it is to be considered
an art or a science’ (1954:50). The problem here, I think, is too
much science and not enough art. Humans are not rabbits, or
fruit fly, or yeast.
Humans use the most productive patches as a high priority
then move to the next most productive and so on? Well,
they might, I suppose, with perfect knowledge and a rather
mechanical view of ‘diet’. Or more likely they won’t, depending
on actual group composition, religious considerations,
food preferences, seasonal variations, local water resources,
shelter availability, relationships with neighbouring groups,
topography, cultural matters, storage technology, gathering
and hunting technology and so on. I don’t think ‘patches’ are
as important as the overall human habitat available, which is
why I thought, and still think, that the slopes and plains of
northern and eastern Australia are a likely initial pathway. The
beaches look good to us, but for a hunter-gatherer without
watercraft I think probably not so much. Sure you can strip all
the shellfish off rocks and dig for them in sand, but then what?
Where is the handy freshwater and shelter that doesn’t face
towards the storms, and topography that allows easy movement
inland? Conversely, these days it is easy to underestimate the
biodiversity and favourable environment of the slopes and
plains, seeing it, as we do now, in its relatively sterile form.
From 1813 on, with the Blue Mountains crossed, these slopes
and plains, so favourable for grazing and later other farming
practices, were subject to rapid, and far greater, change than
elsewhere in the country. Trees were cleared enthusiastically to
expand grazing grass, the hooves of sheep and cattle trampled
the soil, weeds took hold, pastures were fertilised and improved,
fences erected, native animals hunted, creeks dammed, rivers
silted up, and foxes, rabbits, sparrows and honey bees arrived.
The largely bare open hills I see from my window are as much
an artefact of British farming as are the similar hills in England
and as unappealing for Aboriginal occupation now as they
would have been appealing 200 years ago. I think both beach
and semi-arid areas, as permanent settlement areas, were only
occupied as population built up in the woodlands of slopes and
plains and the excess had to spread out further.
This brings me to population. As Allen and O’Connell point
out, there is no evidence for the Birdsell-type calculation that
sees population growing very rapidly from Adam and Eve to a
million in no time at all as it grows exponentially through the
generations: One, two, many … like yeast in a test tube. But it is
human beings involved here, living in groups, with social, cultural,
spiritual, historic, psychological, ecological, physiological and
22 Number 74, June 2012
Forum
structural limitations and advantages, not to mention the effects
of natural disasters. I don’t think any human groups are likely to
increase at the maximum possible rate, nor do I think Aboriginal
groups in different parts of Australia are all going to increase at
the same rate.
Finally, another problem the authors are well aware of.
Projecting our current image of what the present-day Australian
environment is like back into even the recent past, let alone
the distant past, is very misleading. Applying the ethnography
of a handful of well-studied Aboriginal groups from this
very atypical period since 1788 even to other contemporary
Aboriginal groups is fraught with danger. Applying it to, say,
European Mesolithic hunters is even more inappropriate. And
it is equally true that applying what we learnt about Yolgnu,
Warlpiri and other groups in the 1960s and 1970s to the
ancestral ‘Aborigines’ of 40,000 and more years ago simply
assumes what needs to be proved. This is a self-fulfilling
prophecy as damaging as that of fire-stick farming.
This paper is a fine contribution to the ‘Origin of the
Australians’ genre and I congratulate the authors. It is another
step along a way which really relies much more on small steps
than big leaps – much like the colonisation of Australia I suppose.
NOW BRINg ME THAT HORIzON
William F. KeeganFlorida Museum of Natural History, University of Florida,
Gainesville FL 32611, USA [email protected]
Archaeologists seem to face far more complications in making
the crossing to Sahul than the people who accomplished this feat
about 50 kya. Radiocarbon dates are now so numerous that the
presence of humans in Sahul at this early date cannot be ignored.
The question is how and why this passage was accomplished. In
answer, O’Connell and Allen present a very logical argument.
Reconstructing the crossing to Sahul focuses initially on the
availability and type of watercraft. I do not see this as a major
problem. Most large rivers dispatch substantial quantities of
floating debris, including large trees (Keegan and Mitchell 1986).
It is hard to imagine that modern humans, even 50 kya, did not
understand buoyancy. It is relatively easy to construct a raft that
can navigate calm waters (and often are flexible in rough seas1).
Sahul may have been discovered by accident, but accident had
nothing to do with its colonisation.
We face similar issues in the Caribbean (Keegan in press;
see Figure 2). The first evidence for humans in the insular
Caribbean comes from the Greater Antilles, with the earliest
dates clustered around 7 kya. The initial colonists were mobile
foragers who exploited chert sources in the islands and may
have been responsible for the extirpation of sloths and manatees.
Unfortunately, all of the sites from this time are lithic scatters with
no faunal remains. This example is relevant to the colonisation of
Sahul because in the Caribbean people with similar practices did
not cross the shortest water gaps. Based solely on geographical
distance we would expect the first colonists to have crossed
from Florida, the Yucatan or Trinidad. Yet the material evidence
indicates that the first colonists, and the second wave, came from
1 see http://balseros.miami.edu/introduction.htm
the ‘Intermediate’ or Isthmo-Colombian area (Keegan in press).
Expeditions to colonise the islands involved crossing an expanse
of the Caribbean Sea by peoples we assume to have limited
maritime technology (Figure 3).
Efforts to explain the development of watercraft that
were substantial enough to make the crossing to Sahul have
emphasised the exploitation of marine resources. O’Connor et
al. (2011) recently used faunal evidence from the Jerimalai site
in East Timor to highlight the exploitation of coastal and pelagic
resources ca 42 kya. I am not convinced that watercraft were
necessary. Intertidal molluscs are on dry land for ca 12 hours
each day, and many reef species can be collected by walking along
the reef at low tide. Second, Scombridae (tunas), which they
designated as pelagic, actually frequent shallow coastal waters as
juveniles. Given the climatic variations described by O’Connell
and Allen, these Scombridae may not be pelagic. Generalised
marine habitat denominations for taxa do not capture the
behavioral ranges of fishes (Keegan 2009). Nevertheless, this
line of reasoning does suggest a motive for the production
of watercraft.
O’Connell and Allen adopt a similar, but more operationalised,
perspective in their application of optimal foraging theory (OFT).
I agree with their reasoning, which enhances our understanding
of why particular patches were targeted. The issue I have is that
OFT focuses on marginal and average return rates, and ignores
total return rates (Keegan 1986). The marginal return rate is
optimised to achieve a goal, which is defined as enough food to
meet quotidian requirements (total returns or demand). If the
pursuit, capture and handling of additional prey will not provide
a consumable yield, then there is no need to invest additional
time in searching for prey.
Furthermore, if mobile foragers were the original affluent
society, then they had an abundance of time. After foraging
goals were met, foragers had to decide what to do in their ‘free
time’. The possible options are too numerous to mention, but
one option was the collection and consumption of non-optimal
comestibles. Most molluscs have an incredibly low marginal
return rate, yet they are found in coastal and inland sites
throughout the world. In my opinion, the consumption of small
molluscs is an expression of wealth. People eat small molluscs
because they have the luxury, and not the need, to do so.
Figure 2 The timing and location of human migrations in the insular Caribbean (Keegan in press).
23Number 74, June 2012
Forum
OFT provides a formal economic structure to the most basic
daily decisions in people’s lives. Foragers should expand diet
breadth during periods of abundance (high total returns from
the highest ranked foods) by experimenting with non-optimal
foods. Such experimentation can be achieved at relatively low
opportunity costs. In contrast, diet breadth should contract
during periods of stress (e.g. the colonisation of new territories)
because the search for, and pursuit of, the highest ranked foods
in the optimal set offers the most reliable diet (measured in
terms of marginal return rates). Thus, the earliest sites should
reflect a focus on the highest ranked items, such as at the Coralie
site (Keegan 2007, 2010).
The distribution of patches and prey types promotes
dispersion. This centrifugal force is countered by the distribution
of people; a centripetal force that promotes mobility for
the purposes of assembly. O’Connell and Allen are right in
highlighting demographic parameters, and I agree that the
colonisation of Sahul was accomplished by ‘several separate
landing parties’ that maintained contact, that at the very least,
allowed for the exchange of mates.
The missing element in their argument is social
organisation. A small group colonising Sahul may have
been demographically viable, but was it socially viable? All
societies have marriage conventions. The rapid movement
into Sahul, followed by an abrupt stop, probably reflects the
accumulation of a social population sufficient to allow the
socially constructed exchange of spouses without requiring
return voyages to Sunda. Social relations established through
the exchange of spouses require reciprocity (e.g. Jordan et al.
2009; Keegan 2010). Colonisation involves the movement of a
collectivity of individuals who are members of multiple social
groups and who require intercourse with their parent groups.
Island colonisation is not unidirectional.
O’Connell and Allen present a strong case for the biological
imperatives that structured the colonisation of Sahul. What is
missing is a consideration of the social imperatives. Seaworthy
vessels facilitated the exploitation of new areas and the
colonisation of Sahul, but their invention was predicated on the
need for social interactions. Humans developed watercraft to
facilitate contact with other humans, and not solely to permit
the exploitation of marine resources.
COMMENT
Richard CosgroveArchaeology Program, La Trobe University, Bundoora VIC 3086,
Australia [email protected]
O’Connell and Allen propose one of the more structured
archaeological colonisation models of the recent past. They
present a theory of colonisation based on the available
palaeoecological and archaeological data from the earliest
part of the Sahul record. They suggest that their model can be
tested and refuted. If so, it is a key element in any judgement
of the predictive power of a model. It is an advance on earlier
attempts, such as Birdsell, Bowdler and Horton, to explain
the settlement process of Sahul. At least O’Connell and Allen
cartographically recognise that New Guinea and Tasmania
were in fact connected to Australia at that time. Further, some
earlier models were largely immunised against testing. For
example, Bowdler (1977) in her coastal colonisation model
observed that many, if not all, of the earliest sites were now
under the ocean. Fortunately, later discoveries of sites like
Bobongara, Matenkupkum and Matembek, that continued to
be tectonically uplifted over the last 40,000 years, beyond the
highest sea-levels, preserve a record of human behaviour that
remain unaffected by submergence. Indeed, late Pleistocene
sites discovered before 1977 like King’s Table, Kisope, Kenniff
Cave and Koonalda Cave, suggested an altogether different
adaptation process to the one offered up by Bowdler (1977).
What is clear, however, from the contemporary Australian
archaeological data is that its range and variety have become
more complex, both in time and space, so the developments
of explanatory models require a more sophisticated approach.
This paper tackles the newer evidence, although the usefulness
of their model is constrained by some inappropriate
scalar comparisons.
Models are abstractions of reality that should specify the
expected range of attributes used at various scales that can
identify the archaeological correlates of human behaviours.
O’Connell and Allen could have better dealt with this issue,
explaining the links between the various scales of the observed
phenomenon and archaeological processes as identified in the
data sets. The use of ethnographic analogy is not a problem in
itself and they admit that these are only guiding principles. Their
assumptions are explicit, such as certain behaviours like ranking
food, obtaining mates (and keeping them), and social interaction,
which almost certainly took place in the past. However, there is
a lack of sufficiently strong analogy to link the different scales of
observation together.
For example, their model is based on small scale
ethnographic observations that assume that some common
hunter-gatherer behaviours are likely to be embedded in the
archaeology. However, for their model to be useful, predictions
must be made at commensurate scales between the observed
and predicted archaeological, ecological, climatic and
environmental phenomena. Identifying these is difficult, since
a much higher resolution chronology from all early sites is
needed for comparisons, not just climatic data. Their model
does not sufficiently discuss its limitations, especially the fact
that not all early sites can be commensurately compared given
Figure 3 Modern Haitian ‘pi pris’ used for coastal fishing (relevant to OFT its name can be translated as ‘small price’). It is similar to rafts that made the 200 km crossing from Cuba to Florida (from the north coast), and the 400 km crossing to the Cayman Islands (from the south) in 1994 (photograph courtesy of Lisabeth Carlson).
24 Number 74, June 2012
Forum
the variable structural properties of each record. It is noteworthy
that a majority – 63% – of their archaeological data come from
rockshelters of various geologies.
This is crucial if, for example, the ‘narrow diet’ hypothesis
of early colonists is to be supported. How ‘narrow’ is ‘narrow’?
Detailing preservation conditions and site similarity are as
important as identifying specific taphonomic differences
that structure the archaeological evidence. Their discussion
of ‘subsistence and diet breadth’ importantly includes plants,
but, because evidence of plant use is rarely found, predictions
about narrow diet breadths are difficult to make. This limitation
compounds the difficulty of testing their model in any
meaningful way so that the essence of a good model is lost.
Some of the conditional elements in the argument, such as
colonisation and climate change, are discussed at both small
(human response time) and large (Marine Isotope Stage) time
scales. The ‘Archaeological Assessment’ section is problematic,
as only the earliest dates are described. Significantly, other
chronological markers of change and associated material
culture need to describe the process by which settlement took
place and can be understood. In other words, what are the range
of archaeological attributes needed to identify the behavioural
responses and how are these to be tested given the scalar
problems of the data?
In their ‘simple scenario’, O’Connell and Allen hypothesise
that human groups were small, highly mobile and focused
on ‘sweet spots’. These foraging locations, once depleted of
high ranked resources, were abandoned, with human groups
moving to the next location rather than to the next lowest
ranked resource in that location. Identifying such processes is
again a scalar problem in the model, since the archaeological
predications must be made at levels of human responses, not
large scale climate cycles.
Further, sweet spots are where it is argued people pulse
out in the good time and retreat in the bad. These sweet spots,
however, will be equally affected by climatic downturns and
become rather sour spots to live, with populations taking a hit.
As is known, it’s the bad times that control population levels.
The point here is that there is a wide range of geographic
variability across the 11.5 million square kilometre Sahul
continent. The variable effect of modern drought over regional
areas of Australia is something that could be expected in the late
Pleistocene but perhaps of higher frequency and magnitude.
O‘Connell and Allen seem to fall into the trap of assuming
similarity across this vast Sahul continent in terms of climatic
impact, although archaeologically it is quite variable during
the LGM. However, as they recognise, it’s the embedded small
scale, ENSO-like perturbations that dictate population levels.
Thus, momentum for post-drought mobility will be linked
to recovery rates of both population and the palaeoecology.
Modelling the predicted archaeological patterns in both space
and time is dictated by whether group relocation or accessing a
lower ranked food can be identified archaeologically elsewhere.
At this point we are back to the problems of scale and the
resolving powers of the different records.
Overall it’s a positive attempt at model building. At least it
has a greater level of detail, specificity and predictive power than
other pan-continental model building attempts of the past.
THE SHIPPINg NEWS
Ian LilleyAboriginal and Torres Strait Islander Studies Unit, The University
of Queensland, Brisbane QLD 4072, Australia [email protected]
Joe Birdsell was a clever chap. He has had some bad press over the
years, but I don’t think it is really deserved. This paper gives us
another reason to dust off Birdsell’s reputation, as O’Connell and
Allen return to a proposition he raised 35 years ago, namely that
colonisation of the continent was effected with boats that were
more sophisticated than those recorded ethnographically. In a
related paper they published with Hawkes in 2010, O’Connell
and Allen drew attention to Birdsell’s deduction, but not here,
even though for other reasons they cite the chapter in which
Birdsell made the point. It may seem inconsequential in the
greater scheme of things, but we in the Birdsell Liberation Front
are always watching ...
On a less flippant note, this paper (and indeed Birdsell’s
original chapter and the writings of others on the topic of
early Australian watercraft) raises a persistent question for us:
the extent to which ethnographic analogy helps or hinders
archaeological analysis and interpretation. The discipline goes
back and forth on this issue without resolution, and as much as
I agree with the thrust of O’Connell and Allen’s present paper, it
seems to me that they want to have it both ways when it comes
to the use of ethnographic examples, without any explanation of
why they think it’s justified to do so.
The accumulating empirical evidence for the use of capable
ocean-going watercraft in and around Sahul 40,000-50,000
years ago may confront our crustier colleagues, but it seems
incontrovertible. This in turn has major implications for the
initial movement of modern people out of Africa around the
Southern Dispersal Arc, and all of it shines a very strong light
on orthodoxies concerning human evolution and progress.
This is truly exciting stuff! Yet if we are turning things on their
heads in this way and, like O’Connell and Allen, explicitly ‘reject
inferences ... that rely on assumptions about marine technology
constrained by the Australian ethnographic record’, we have to
ask why we still rely on ethnographic analogy as heavily as they
do in other parts of this paper. If the boats were different back
then, why assume that what was true of ethnographic subsistence
and settlement patterns ‘was also true for early Sahul colonisers’?
Why assume ‘foraging options ... were similar to those reported
historically’? And so on, whether regarding women’s roles,
population growth and density, or diet breadth.
I know we have to start somewhere, analytically-speaking,
and O’Connell and Allen have certainly been clear about the
assumptions they are making in this regard. As a discipline
we’ve been insisting for a while now, though, that the past is
a foreign country where they did things differently. It seems
inconsistent to be so forceful about boats in this regard, but to
make no comment about how appropriate it might be to use
analogies in the other instances. This is especially true of matters
concerning subsistence and settlement if, as O’Connell and Allen
contend, ‘human populations ... remained much smaller than
sometimes imagined’, and thus presumably also much smaller
than the ethnographically-recorded (or at least guesstimated)
populations whose foraging and habitation patterns underpin
25Number 74, June 2012
Forum
the authors’ optimal foraging approach. The authors themselves
seem to confirm this point when they note that ‘Historically
known patterns of economy, technology and demography
emerged only in the Holocene’.
Where to from here? I understand, and am philosophically
inclined towards, the critique of inductive reasoning O’Connell
and Allen made with Hawkes in 2010, which is behind
their willingness to use ethnographic analogy in deductive
theoretical models. As their own work demonstrates in this
paper, though, it is induction from facts emerging from the
ground in places such as Timor and the Bismarck Archipelago
that is forcing the issue with boats. This is not the place for
extended methodological discussion, but clearly we need
to move beyond what are now very stale arguments about
deductive vs inductive reasoning. Quality scholarship usually
entails a mix of both, in proportion to the nature and quality
of one’s empirical data, the internal logical consistency of
one’s deductive theoretical models, and the fit between one’s
data and one’s models. Despite their seemingly hard-nosed
rhetoric, this is pretty much what O’Connell and Allen do
in this paper. Deductive models can be based on almost any
sort of information, including potentially quite misleading
ethnographic analogy. One needs to make one’s sources of
such information clear. This the authors do. One should also
make explicit the limitations of that information, in this case
ethnographic analogy, which they don’t (unless it’s about
boats). I may have become a bit obsessive about it, but in the
end the omission really detracted from my reading of the paper.
dANCINg ON PINS: TENSION BETWEEN CLEVER THEORy ANd MATERIAL RECORdS IN AUSTRALIAN ARCHAEOLOgy
Peter HiscockSchool of Archaeology and Anthropology, The Australian
National University, Canberra ACT 0200, Australia
Thomas Aquinas (ca 1225-1274 AD), like many medieval
philosophers, explored the nature of his world through detailed
theorising – building models through logic and dialectic. This
model building was focused on the consequences of phenomena
believed to exist, with an emphasis on the exercise of reason
but not of empiricism. Consequently, although Aquinas was a
crucial contributor to modern thinking, his theorising involved
questions as abstruse as whether there was defecation in Paradise
or whether angels were material entities (often misquoted as how
many angels could dance on the point of a pin). Those medieval
debates displayed impressive intellectual dexterity, but in
retrospect we can view the sophisticated thought as squandered
on illusory concerns and fantastical visions of the world.
Comparable situations occur repeatedly in science, because
diagnosing what happens in the world can be extremely difficult.
Perhaps, as we archaeologists always claim, it is more difficult
to interpret the human past from material residues than to
do many other kinds of science. In any case, tensions between
explanatory theory and the inferred pasts that are the subject
of those explanations have occurred throughout the history
of archaeology. The issue is not that explanatory models are
eventually refuted but that, at the time they were proposed, they
sought to explain impressions of the past which were empirically
dubious or even demonstrably incorrect. A well-known instance
of this was Bowdler’s (1977) coastal colonisation model, which
at the time it was formulated was already contradicted by inland
sites such as Kenniff Cave. Currently, the same tensions are
displayed in the extended debates over whether the temporal
coincidence between the arrival of humans and the extinction
of megafauna is evidence of human culpability, despite the
uncertainty about the timing of either arrivals or extinctions
and the existing inability to adequately test human contributions
(see Field and Wroe 2012). Of course these tensions can be
extraordinarily productive, driving scholars to collect further
data and to reconceptualise their analytical framework, but
creative outcomes are most likely when scholars acknowledge and
engage with the uncertainty and interpretative disagreements.
In this interesting paper O’Connell and Allen offer
explanations that are both elaborate and novel for past cultural
events and processes. The use of principles from behavioural
ecology, and specifically OFT, is intriguing and clearly
worthwhile. These approaches to understanding behavioural
decisions are known to be powerful and will have intuitive appeal
to researchers in the field. It is therefore of considerable value
that they have chosen to explore the colonisation of Sahul in this
way. My concern here is not with the application of ecological
perspectives to the range-expansion of humans, since that
seems both appropriate and inevitable, but to the question of
what is being explained. In presenting their analysis they discuss
archaeological data that supports a particular image of the
colonisation process, specifically a late, rapid dispersion of Homo
sapiens from Africa to Australia, scattering in low population
densities, using only simple technologies across the continent
and being subsequently largely isolated. Such an image is, and
should be, contested. Certainly O’Connell and Allen (2004) have
given their reasons for advocating a late date for colonisation, but
the reconstructions of the colonising event by other researchers
do not accord with theirs, and the evidence is clearly arguable.
Certainly many of the components of the colonisation story told
by O’Connell and Allen should be considered carefully. I alert
readers to some of the obvious claims they may wish to consider.
The Colonisation of Sahul Occurred 44-46 kya? This assertion ignores sites that have older dates claimed to be
in association with cultural material, and fails to account for
the sampling issues in investigating the archaeological record.
Furthermore, it does not reflect the dating uncertainties of
sites such as Parkupirti which may be older than 40 kya, but
how much older is currently very poorly defined. Additionally,
I and others have repeatedly made the case that the oldest sites
found, especially in the south and inland of the palaeocoast,
are not likely to date the arrival of humans but more likely
reveal the time when population and landscape use had
risen to the point of being archaeologically visible. So even if
claims for earlier sites are eventually rejected and O’Connell
and Allen are supported in their assertion that uncontested
dates for Homo sapiens fall in a relatively narrow time range
(44-46 ky BP), this must be considered a time when humans
have established themselves across Australia, not necessarily
26 Number 74, June 2012
Forum
the time at which colonisation occurred. Consequently,
assertions about the timing and rapidity of movement and
the relationship of that movement to sea-level movement are
speculative at best.
Occupation in the Early Millennia Involved Infrequent, Short-Term Occupation by Small, Highly Mobile Human Groups?This oft-cited claim is usually based, as it is in O’Connell and
Allen’s paper, on the discovery of sites that are individually small
in area and/or low in perceived artefact discard rates. While this
might be a consequence of small groups staying only for short
times this literal interpretation is not without its uncertainties.
Critical questions remain largely unanswered, such as what
proportion of early occupation took place inside rockshelter sites,
what proportion of early occupation sites are preserved, and
what selection of archaeological sites have been recovered (see
Langley et al. 2011). At any time period the interpretation of site
numbers and sizes in terms of population and residential group
size and mobility is tricky (see Attenbrow 2004), and for the
earliest period of human occupation I suspect more discussion
will be required to resolve what interpretations are sustainable.
Early Sahul Technology was Simple and Displayed Limited Variation through Time and Space? It is unfortunate that the authors do not elaborate and
substantiate this statement, for it is contrary to my impressions
of the evidence (see arguments in Hiscock 2008). Not only are
the technologies revealed in early assemblages of flaked stone
artefacts quite diverse across the continent, we now have evidence
of bone and shell working creating tools/ornaments and of the
production of edge-ground axes at or before about 40 kya. Using
the late colonisation chronology advocated by O’Connell and
Allen these data indicate a geographical and temporal diversity
of technologies immediately after the dispersal of humans
across the continent. Additionally the notion of simplicity is
undemonstrated, and relies on out-dated depictions of stone
assemblages that have long been challenged. O’Connell and
Allen conclude their paper with an admission that the pelagic
fishing, and the elaborate gear that probably supported it, is
hard to explain in their model. Indeed that does merit further
investigation, because it is one possible pointer to a colonisation-
era cultural system in which population might have been greater
and empty territory often less available than their model implies.
These are just a few of the inferences about the past that
O’Connell and Allen present and attempt to explain. Each
is debatable and warrants investigation, because, while the
ecological approach applied in the paper is intriguing, if it
is applied to visions of the past that are more fantastical than
evident, it too may be squandered.
BOATS, STONES ANd RISKS
Robin TorrenceAustralian Museum, 6 College Street, Sydney NSW 2010,
Australia [email protected]
This interesting model combines evolutionary ecology with the
prime mover of population pressure invoked to kick off the
human colonisation of Sahul. I question whether population
pressure was necessary to ‘push’ groups only targeting resources
with high returns. More likely, once they left Africa, colonising
populations would have been ‘pulled’ by the high ranked
resources in adjacent unoccupied land and seascapes. Given the
nature of tropical environments, especially in Sahul, such food
sources tend to be relatively sedentary (e.g. plants, molluscs,
varanids and arboreal mammals). Their rapid depletion would
necessitate frequent moves, which may help explain the use of
boats, especially because these could reduce the stress of constant
mobility on women and children. Marine technology would
be better incorporated into the model by acknowledging its
initial role as transport required by a subsistence focus on high
return resources, rather than for fishing, which is perhaps better
conceived as a by-product.
Continuing the theme of technology, the authors rightly
attack northern latitude scholars, such as Mellars, who continue
to make simplistic, direct links between the evolution of human
intelligence and the nature of stone tool assemblages, despite
the contradictory archaeological record of Homo sapiens in
Sahul which is dominated by a simple flake-core technology (cf.
Brumm and Moore 2005). Instead, O’Connell and Allen try to
show that a better way to explain human behavior is to study
the relationships between ‘the ecological context and foragers’
intellectual solutions to the problems of making a living’.
Whereas I agree that the particular configuration of stone
technology is the consequence of people finding optimal
solutions to problems posed by both their physical and social
environments (Torrence 1989), I run into difficulties with the
specific hypotheses they put forward, although I appreciate these
are preliminary attempts to exemplify their proposed approach.
Accurate definitions underpin all good modelling. Tools
and technology refer to very different entities. Technology
encompasses ‘physical setting, social context, actors, knowledge,
energy sources, raw materials, tools, actions and outcomes’
(Torrence 2001:74; cf. Bleed 1997). Boats are not necessarily
‘complex’ tools, but the knowledge and skills required to navigate
across the ocean may be very sophisticated. Does this mean that
pelagic fishing is necessarily more complex than the behaviour
and knowledge associated with successfully hunting kangaroos?
When O’Connell and Allen ask ‘why is Pleistocene Sahul
technology simple in some settings but complex in others’ , they
are actually mixing up stone or shell tools with fishing technology.
For the sake of having something concrete to observe,
archaeologists focus on the tools themselves rather than the
technology. Considering pursuit tools, the opposite of what
O’Connell and Allen predict is well-known. For recent hunter-
gatherer tool-kits, there is a strong negative correlation between
diet breadth and the complexity of both individual tools and
assemblages (Bamforth and Bleed 1997; Torrence 1989). The
reason for this pattern is that a wide range of complex and reliable
tools are required in situations where there is a high probability
of failing to acquire adequate resources. High risk is often, but
not always, correlated with specialised diets: environmental
parameters, such as seasonality, are also important. Where a
broad range of resources are targeted, risk of failure is generally
low and less time and energy are invested in tools. In these
situations generalised tools further reduce costs. I suspect the
same is true for processing and/or handling tools. Where the
27Number 74, June 2012
Forum
risk of failing to acquire sufficient resources is high, such as with
specialised diets, complex tools are required.
A further problem with the modelling in this article is
that the majority of tools used directly in subsistence tasks
during the late Pleistocene were made from plant materials
rather than stone. The primary role of flaked stone was for
manufacturing these tools rather than in pursuit or handling
of resources (stone points are only common later in time).
Since manufacture can easily be embedded in other activities,
the risk of failing to have the necessary tools prepared in
the required timeframe would have been low. Consequently,
the stone tools for producing other tools were fairly simple,
although we need better use-wear studies to investigate if and
how they were hafted.
Given the scarcity of evidence, I question O’Connell and
Allen’s emphasis on pelagic fishing, especially as a planned
subsistence activity, rather than as an adjunct to travel. If boat
technology was created and maintained primarily for voyaging,
then dropping a hook and line over the side would not have been
a costly or risky undertaking and so the unpredictable returns
would have been inconsequential. Given the paucity in late
Pleistocene contexts of fish remains and associated tools, it is
difficult to be certain about the relative contribution of pelagic
fishing to the diet, but I suspect that boats were primarily
designed for transport and exploration, and not for direct
capture of resources.
This brings up a major question raised but not answered
in the paper: despite the lure of high ranked resources lurking
just over the horizon, why did people risk deep-sea journeys?
Modelling the costs of search and pursuit takes us a long way
toward predicting human subsistence and consequences for
colonisation, but a more complete picture requires understanding
the factors behind taking and managing risk.
COLONISINg SAHUL
Mark W. MooreArchaeology and Palaeoanthropology, University of New
England, Armidale NSW 2351, Australia [email protected]
A vague notion of ‘wanderlust’ seems to be the driving force in
many narratives about hominin migration (e.g. Dennell and
Roebroeks 2005), but, true to the zeitgeist, O’Connell and Allen
have shown us that wanderlust is all about food. The strength
of behavioural ecology is the explicit nature of the underlying
assumptions and the clear connection between forager theory,
predictive statements and archaeological evidence. Summarising
several optimal foraging models, O’Connell and Allen conclude
that optimising hominins are pulled from patch to patch by the
serial depletion of highest ranked resources. The logic of their
scenario is straightforward: the archaeological record shows
that humans colonised Wallacea and Sahul, and the theoretical
model stipulates that forager movement is linked to exploitation
of highest ranked prey, therefore colonisation was driven by the
pursuit of highest ranked prey. One might question whether
certain assumptions of optimal foraging models – for example,
that foragers have perfect resource knowledge and the perfect
ability to exploit it – would apply to the first wave of colonists
to cross the Wallace Line, but the successful colonisation itself
might be de facto evidence that the costs of imperfect knowledge
were not prohibitively high. O’Connell and Allen posit that,
after colonisation, movement between patches in pursuit of
highest ranked prey became the norm as foragers made nearly-
continuous readjustment to unstable climatic conditions.
Issues of colonisation rates and routes aside, the model
succeeds or fails by the prediction that the Pleistocene menu
was dominated by highest ranked prey. The data reviewed
by O’Connell and Allen are sketchy and, in most cases, the
regional and local subsistence resources surrounding early
archaeological sites are not sufficiently understood to warrant
informed scenarios of resource ranking (for instance, where are
the Australian megafauna?). Perhaps in light of this limitation,
O’Connell and Allen nod to the established tradition of
continent-wide generalisation and essentially treat all of Sahul
as if it were one patch. The continental zones with the greatest
density of highest ranked resources are termed ‘sweet spots’ and,
consistent with the model, the empirical evidence shows that
most early archaeological sites in Sahul are comfortably nestled
within them. However, it is not clear that the empirical pattern
identified by O’Connell and Allen is distinct from patterns
that might be predicted from the unspecified prior narratives
they term ‘minimalist’. The issue here is scale, and while the
data marshalled by O’Connell and Allen seem to support their
scenario, the true test will turn on detailed regional and site
specific reconstructions of palaeoclimate and subsistence.
O’Connell and Allen conclude their discussion with the
generalisation that simple technologies should correlate with
diets focused on highly ranked prey because greater investments
in finding food outweigh technological improvements in
obtaining it. This is a surprising observation, as many complex
forager technologies – particularly in North America and
Eurasia – are associated with the pursuit of highly ranked prey
species. Since a technological approach is an unarguable aspect
of hominin adaptation through the ‘extended phenotype’ (after
Boone and Smith 1998), a key challenge is to differentiate the
attributes of a technology that directly improve reproductive
fitness from those that are, at best, only indirectly related to
adaptation – what Bettinger (1991) has characterised as ‘adaptive’
and ‘adjustive’ traits, respectively.
For example, evolutionary models of flaked stone
technology implicitly assume that the tool as a whole was
adaptive and that inadequate tools directly compromised
fitness (Surovell 2009; Torrence 1986). But one might infer
from O’Connell and Allen that the only truly adaptive part
of a stone tool was the sharp edge, and the shape of the stone
bearing that edge – and thus the nature of the reduction
sequence that created the shape – was unimportant for
reproductive fitness, at least in the pursuit of highly ranked
prey. Most evolutionary narratives suggest that our early
hominin ancestors improved their fitness by creating sharp
edges for food procurement through least-effort flaking
(Ambrose 2001). This approach was exclusive among pre-
modern hominins for at least 1 my. The concern with tool
shapes that subsequently emerged among some (but not all)
modern human foragers was perhaps only an ‘adjustment’
within the extended phenotype and was only indirectly
related, or perhaps unrelated, to hominin fitness; direct fitness
requirements continued to be met through simple, least-
28 Number 74, June 2012
Forum
effort edge production similar to that seen among the modern
human colonisers of Sahul. Indeed, most complex stone
technologies worldwide, associated with a variety of foraging
strategies, included a least-effort reduction trajectory (often
under-analysed by archaeologists and sometimes ignored
altogether) alongside more complex ‘add-on’ approaches
to modifying stone shapes (e.g. Moore et al. 2009). Indeed,
once variables such as manufacturing skill, time investments
and ancillary costs of hafting are considered, it is rarely self-
evident that tools made by complex approaches significantly
improved the handling efficiency of the tools they augmented
(e.g. backed microliths vs flakes; tula adzes vs retouched
flakes, stone-tipped spears vs single-piece wooden ones) (cf.
White 1977). Perhaps the development of flaked stone tool
complexity beyond least-effort stone flaking had negligible
effects on the choices directly modelled by cultural ecologists,
and instead reflected adjustments to other parts of the
foragers’ extended phenotype.
ANd THE ANSWER IS 42 …
Judith FieldSchool of Biological, Earth and Environmental
Sciences, University of NSW, NSW 2052, Australia
The final line in the Restaurant at the End of the Universe quote
is a brilliant beginning to this interesting and ambitious paper
by O’Connell and Allen and speaks of the likely explanations
for the initial forays into, and expansion across, Sahul. That
‘42’ may be the answer to the meaning-of-life question, also
from this series, and perhaps the answer to the question
of human arrivals in Sahul – tells us something we already
know – that the fun is in the journey. I suspect all aspects of
this model will be debated and tested for a long time to come,
but the beauty of it lies in its breadth and the challenges that
will ‘feed’ a whole generation of archaeologists to come.
Having been engaged for some time in the ‘seemingly
eternal’ and faith-based debate about megafauna and its
relevance in the archaeological sphere, there are some issues
I would like to expand on that may contribute to the discussion.
O’Connell and Allen argue that one of the drivers for rapid
expansion was ‘serial depletion of high ranked prey’, which
on face value would support (and possibly revive) Flannery
and colleagues’ approach to the extinction process. The Steve
Wroe-coined phrase comes to mind – ‘spear-wielding hoards
hacking their way through startled megafauna’ – but this time
let’s throw in some of those extant species. If people arrived
on the northwest coast of the continent, the pickings may
have been sparse indeed and this alone would have accelerated
movement through the region. Entering via the Birds Head of
New Guinea may have been a different proposition altogether.
While the suite of animals and plants that were encountered
were entirely new, the ‘package’ of knowledge brought with
them may have enabled small populations to quickly adapt to
the new landscapes they encountered (as argued by Denham
et al. 2009). Certainly the discovery of Pandanus and yam in
the early occupation horizons from the Ivane Valley in the
eastern highlands of New Guinea points to this capacity
(Summerhayes et al. 2010). The observations by Geoff Hope
on Cyathea above the tree-line in this region also speaks of
important carbohydrates being readily accessible to, if not
exploited by, people in the recent past.
As always it is the preservational limitations of Australian
fossil sequences that inhibits a more fruitful exploration of
the ideas concerning exploitation of many animal species
in the Pleistocene record. What has become clear in the last
few years is the lack of evidence for a human-megafauna
association, despite the limited evidence for a temporal
overlap. Our scrutiny of the fossil record places only a handful
of megafauna still wandering around the landscape post-50 ka.
The spatial distribution of these remnant megafauna is
unclear and does not seem to be continuous across the
landscape. Yet this has not inhibited the supporters of human-
driven extinctions to push their particular barrow. The most
recent case is the report from Lynch’s Crater (Rule et al. 2012)
where they argue, on the basis of evidence from a sediment
core (analysing Sporormiella spores, charcoal and summary
diagrams of scleropyllous vegetation), that the extinction of
megafauna at 41 ka (where this date comes from in their paper
is unclear), drove changes in vegetation composition because
of ‘relaxed pressure’ on vegetation. Sporormiella spores are
used as a proxy for megafauna (but see critique by Feranec et
al. 2011). Despite these largely unproven assertions about the
timing of megafaunal demise, and the dismissal by Rule et al.
of climate change as an important factor, it now seems clear
to many that megafaunal extinctions cannot be considered an
archaeological question. For the most part, megafauna played
no significant role in the success or otherwise of colonising
populations in Sahul (Field and Wroe 2012).
I believe one of the important points that emerges from
this paper is the suggestion that we should broaden our gaze
when it comes to defining the particular conditions that
allowed/facilitated the movement of people to Sahul for the
first time. While to some extent we are still constrained by the
current sequences and dates for the earliest sites, O’Connell
and Allen challenge us to move away from the ‘lowest sea-
level moment’ when looking at the first arrivals. We are in
no doubt that the first Australians were behaviorally modern
and, as such, modes and routes of travel may have been
more complex than we can currently predict. One point that
has received little attention is the importance of access to
freshwater, a point raised by Horton (1984) some time ago as
it related to the faunal extinction process. It would seem that
the same principle may apply to colonising humans, especially
on the Australian continent. It is a fact that MIS-3 was not the
relatively stable climatic period as asserted by some authors
(e.g. Miller et al. 2005 and others), and clearly evident from
the Antarctic ice cores (EPICA Community Members 2006).
This instability must have been a determining factor in the
presence or absence of free water in the more arid regions
and also in the timing of movements across the Australian
continent. The unpredictability of sudden swings in climate
may have meant that the earliest archaeological records
reflected the necessarily ephemeral nature of occupation
across most landscapes.
So much to think about – this paper has given us a compass:
I look forward to seeing how it is used.
29Number 74, June 2012
Forum
RESPONSE
This paper developed as a Forum piece at the suggestion of
previous editor Sean Ulm, with continued positive input from
Lynley Wallis and Heather Burke. We thank them, and all the
contributors who took the time to respond.
Codding and Bliege Bird add an important dimension to
our argument by distinguishing the spatial (and by implication
temporal) scales at which foraging decisions operate – among
resources within patches, between patches within habitats, and
between habitats on continental and sub-continental scales.
Davidson, Dillehay, Horton and Keegan note that our model
makes no reference to considerations other than those related to
subsistence that might constrain movement into and across Sahul.
Our reductionist approach is intended to develop falsifiable
propositions about human behaviour in the distant past – it is
deliberately simplistic. It ignores social (except gender) and other
potential complexities, not because they were unimportant, but
because sidestepping them for the moment simplifies hypothesis
testing. We do not expect subsistence-related decisions to predict
or explain all aspects of past behaviour. Instead, we seek to
determine how much of that behaviour can be anticipated in
those terms, and to highlight what is not accountable in this way.
The exercise is intended to be iterative – we simply picked the
easiest, theoretically and empirically best-grounded place to start.
We recognise and share Cosgrove’s concerns about reconciling
data on different scales. To maintain simplicity and conserve
space this was not the place to pursue them. We use Oxygen
Isotope Stages and the like as familiar conveniences, but, with
Cosgrove, we believe that shorter-term climatic oscillations, like
ENSO – recognisable in human life spans – are the significant
modifiers of foraging behaviour.
Davidson and others note that the model assumes what
ecologists refer to as ‘complete’ knowledge of foraging
opportunities, and the costs and benefits of exploiting them.
The forager’s task of accumulating this knowledge is probably
less daunting than they imagine, especially at the resource and
patch scales on which short-term decision-making operates,
and especially with humans who can exchange knowledge
through language. Foragers’ abilities to assess the costs and
benefits of immediate subsistence-related opportunities and act
on them optimally is clearly indicated by the past four decades
of ethnographic research conducted within the framework of
behavioural ecology in Africa, the Americas and Australia. That
same research underlines foragers’ capacity to adjust to changes in
those constraints in real time. Davidson’s reference to Alyawarra
women’s use of O’Connell’s vehicle to reach and exploit high
ranked resources (mainly geophytes, never low ranked seeds)
in distant patches illustrates that ability. In archaeological
perspective, such adjustments, including those involving
previously unfamiliar resources, should be ‘instantaneous’. All
else being equal, we expect delays in the exploitation of new
opportunities to reflect ongoing assessments of economic and
ecological trade-offs, not slow learning. Tree and grass seeds
enter Australian diets late not because of a long-term lack of
relevant knowledge, but because the costs of exploiting them are
high relative to those associated with the use of many other foods.
Lilley chides us for what he sees as an inconsistent appeal
to ethnographic information. In our view, the question isn’t
whether to appeal to it, but how. Direct-historical analogies of
the kind once popular among Sahul archaeologists are now
often seen to be inappropriate, especially in late Pleistocene
settings. But rejecting any reference to human behaviour in the
present makes the archaeological enterprise entirely descriptive
and chronological. Simply accumulating more archaeological
data and ‘discussing’ them, the gambit Hiscock advocates
here and elsewhere, is a retreat to mindless empiricism and
baseless speculation, a Luddite strategy that sharply restricts
archaeology’s potential to reconstruct and explain the human
past. Our approach appeals to a uniformitarian theoretical
framework, capable of generating testable predictions about
many aspects of behaviour and shown empirically to be
applicable across a wide range of organisms, humans and fruit
flies alike (pace Horton). Among its most important virtues in
this context is its ability to generate predictions about patterns
in human behaviour unrepresented in ethnography. Codding
and colleagues’ (2011) recent work on the sexual division
of labour, specifically on the circumstances under which
men’s prey choices might be expected to trump women’s as
a determinant of residential site location in middle and low
latitude habitats, is a good example and one that we expect to
build on in future work.
On the relationship between technological complexity and
diet breadth: our sense of the global record, especially across the
late Pleistocene and Holocene, differs from the views expressed
by Torrence, Dillehay and, to some degree, Moore. Patterns in
Australian lithics are clearly consistent with our position, as are
those in, for example, western North American Paleoindian
vs Archaic contexts. Broader review is beyond us here, but
this difference of opinion may be at least in part a matter of
definition. For reasons stated in our paper and developed at
length in the sources we cite, we expect greater diet breadth to
be associated with increased complexity in handling technology,
defined as that connected with the post-encounter procurement
of prey, including all tools used in the pursuit, capture and
processing of those resources. We will explore this issue at greater
length elsewhere.
Minor points: (1) Our previous arguments for developed
watercraft did not rely heavily on the presence of pelagic fish
bones or fishing tackle in pre-LGM sites, but rather on the
need to maintain biologically viable colonising groups across
large distances. Nevertheless, the availability of such craft
would have facilitated pelagic fishing as a consequence; (2)
Our model is not simply an appeal to population pressure
as a catalyst for movement, but instead draws attention to
the complex interplay between unevenly spread populations
and resources, and the choices made about those resources
in relation to such movement; and, (3) Despite frequent
appeal to risk avoidance as a determinant of pre-European
technological diversity across Australia, the link between
the two remains poorly developed (White 2011). Recent
research from the perspective of behavioural ecology clearly
indicates that the significance of risk varies across habitats
and economies and (importantly) by gender (e.g. Hawkes and
Bliege Bird 2002). We look forward to further debate on this
issue. For the moment we side with Field and think the answer
to this and wider questions might still be ‘42’. We trust that
this piece opens a few doors.
30 Number 74, June 2012
Forum
ReferencesAmbrose, S.H. 2001 Paleolithic technology and human evolution. Science 291:1748-
1753.
Attenbrow, V. 2004 What’s Changing: Population Size or Land-Use Patterns? The
Archaeology of Upper Mangrove Creek, Sydney Basin. Terra Australis 21.
Canberra: Pandanus Books.
Bamforth, D. and P. Bleed 1997 Technology, flaked stone technology and
risk. In C. Barton and B. Clark (eds), Rediscovering Darwin: Evolutionary
Theory in Archaeological Explanation, pp.109-140. Archaeological Papers
of the American Anthropological Association 7. Arlington: American
Anthropological Association.
Beck, W. 1992 Aboriginal preparation of Cycas seeds in Australia. Economic Botany
46(2):133-147.
Bettinger, R.L. 1991 Hunter-Gatherers: Archaeological and Evolutionary Theory.
New York: Plenum Press.
Birdsell, J.B. 1953 Some environmental and cultural factors influencing the structuring
of Australian Aboriginal populations. American Naturalist 87:171-207.
Boone, J.L. and E.A. Smith 1998 Is it evolution yet? A critique of evolutionary
archaeology. Current Anthropology 39:S141-S173.
Bowdler, S. 1977 The coastal colonisation of Australia. In J. Allen, J. Golson and R.
Jones (eds), Sunda and Sahul: Prehistoric Studies in Southeast Asia, Melanesia
and Australia, pp.205-246. London: Academic Press.
Boyd, R. and P.J. Richerson 1987 The evolution of ethnic markers. Cultural
Anthropology 2(1):65-79.
Brumm, A. and M. Moore 2005 Symbolic revolutions and the Australian
archaeological record. Cambridge Archaeological Journal 15:157-175.
Bulbeck, D. 2007 Where river meets sea: A parsimonious model for Homo sapiens’
colonisation of the Indian Ocean rim and Sahul. Current Anthropology 48:315-321.
Charnov, E.L. 1976 Optimal foraging, the marginal value theorem. Theoretical
Population Biology 9:129-136.
Codding, B.F., R. Bliege Bird and D.W. Bird 2011 Provisioning offspring and others:
Risk-energy trade-offs and gender differences in hunter-gatherer foraging
strategies. Proceedings of the Royal Society B 278:2502-2509.
Davidson, I. 1992 Archaeologists and Aborigines. The Australian Journal of
Anthropology 2(2):247-258.
Davidson, I. in press Hunter-gatherers in Australia. In V. Cummings, P. Jordan and
M. Zvelebil (eds), Oxford Handbook of the Archaeology and Anthropology of
Hunter Gatherers. Oxford: Oxford University Press.
Denham, T., R. Fullagar and L. Head 2009 Plant exploitation on Sahul: From
colonisation to the emergence of regional specialisation during the Holocene.
Quaternary International 202:29-40.
Dennell, R. and W. Roebroeks 2005 An Asian perspective on early human dispersal
from Africa. Nature 438:1099-1104.
Dillehay, T.D. 2011 From Foraging to Farming in the Andes: New Perspectives on
Food Production and Social Organization. London: Cambridge University Press.
EPICA Community Members 2006 One-to-one coupling of glacial climate
variability in Greenland and Antarctica. Nature 444:195-198.
Erlandson, J.M. 2001 The archaeology of aquatic adaptations: Paradigms for a new
millennium. Journal of Archaeological Research 9:287-350.
Erlandson, J.M. 2010 Neptune’s children: The origins and evolution of seafaring. In
A. Anderson, J. Barrett and K. Boyle (eds), Global Origins and Development of
Seafaring, pp.18-27. Cambridge: McDonald Institute for Archaeological Research.
Erlandson, J.M., M.H. Graham, B.J. Bourque, D. Corbett, J.A. Estes and R.S. Steneck
2007 The Kelp Highway hypothesis: Marine ecology, the coastal migration
theory and the peopling of the Americas. Journal of Island and Coastal
Archaeology 2:161-174.
Feranec, R.S., N.G. Miller, J.C. Lothrop and R.W. Graham 2011 The Sporormiella
proxy and end-Pleistocene megafaunal extinction: A perspective. Quaternary
International 245:333-338.
Field, J. and S. Wroe 2012 Aridity, faunal adaptations and Australian late Pleistocene
extinctions. World Archaeology 44:56-74.
Fretwell, S.D. and H.L. Lucas 1969 On territorial behavior and other factors
influencing habitat distribution in birds I. Theoretical development. Acta
Biotheoretica 19:16-36.
Golson, J. 1971 Australian Aboriginal food plants. In D.J. Mulvaney and J. Golson
(eds), Aboriginal Man and Environment in Australia, pp.196-238. Canberra:
Australian National University Press.
Hawkes, K. and R. Bliege Bird 2002 Showing off, handicap signalling and the
evolution of men’s work. Evolutionary Anthropology 11:58-67.
Hiscock, P. 2008 Archaeology of Ancient Australia. Routledge: New York.
Horton, D.R. 1984. Red kangaroos: Last of the Australian megafauna. In P.S. Martin
and R.G. Klein (eds), Quaternary Extinctions: A Prehistoric Revolution, pp.639-
680. Tucson: The University of Arizona Press.
Jordan, F.O., R.D. Gray, S.J. Greenhill and R. Mace 2009 Matrilocal residence
is ancestral in Austronesian societies. Proceedings of the Royal Society B
276(1664):1957-1964.
Keegan, W.F. 1986 The optimal foraging analysis of horticultural production.
American Anthropologist 88:92-107.
Keegan, W.F. 2007 Taino Indian Myth and Practice: The Arrival of the Stranger King.
Gainesville: University Presses of Florida.
Keegan, W.F. 2009 The synergism of biology and culture. Journal of Island and
Coastal Archaeology 4:240-248.
Keegan, W.F. 2010 Demographic imperatives for island colonists. In A. Anderson, J.H.
Barrett and K.V. Boyle (eds), The Global Origins and Development of Seafaring,
pp.171-179. Cambridge: McDonald Institute for Archaeological Research.
Keegan, W.F. in press Caribbean: Colonisation and subsequent population
movements. In P. Bellwood (ed.), Global History of Human Migrations, Vol. 1.
New York: Wiley-Blackwell.
Keegan, W.F. and S.W. Mitchell 1986 Possible allochthonous Lucayan Arawak artifact
distributions, Bahama Islands. Journal of Field Archaeology 13:253-258.
Langley, M.C., C. Clarkson and S. Ulm 2011 From small holes to grand narratives:
The impact of taphonomy and sample size on the modernity debate in
Australia and New Guinea. Journal of Human Evolution 61:197-208.
McGrew, W.C. 1987 Tools to get food. Journal of Anthropological Research 43:247-
258.
Miller, G.H., M.L. Fogel, J.W. Magee, M.K. Gagan, S.J. Clarke and B.J. Johnson 2005
Ecosystem collapse in Pleistocene Australia and a human role in megafaunal
extinctions. Science 309:287-290.
Moore, M.W., T. Sutikna, Jatmiko, M. Morwood and A. Brumm 2009 Continuities
in stone flaking technology at Liang Bua, Flores, Indonesia. Journal of Human
Evolution 57:503-526.
Morwood, M.J., D.R. Hobbs and D.M. Price 1995 Excavations at Sandy Creek 1 and
2. In M.J. Morwood and D. Hobbs (eds), Quinkan Prehistory, pp.71-91. Tempus
3. St Lucia: Anthropology Museum, The University of Queensland.
O’Connell, J.F. and J. Allen 2004 Dating the colonisation of Sahul (Pleistocene
Australia-New Guinea): A review of recent research. Journal of Archaeological
Science 31:835-853.
O’Connell, J.F. and K. Hawkes 1981 Alyawara plant use and optimal foraging theory.
In B. Winterhalder and E.A. Smith (eds), Hunter-Gatherer Foraging Strategies,
pp.99-125. London: University of Chicago Press.
O’Connell, J.F., P.K. Latz and P. Barnett 1983 Traditional and modern plant use
among the Alyawarra of Central Australia. Economic Botany 37(1):80-109.
O’Connor, S., R. Ono and C. Clarkson 2011 Pelagic fishing at 42,000 years before
the present and the maritime skills of modern humans. Science 334:1117-1121.
Oswalt, W.H. 1976 An Anthropological Analysis of Food-Getting Technology. New
York: John Wiley.
Pope, K.O. and J.E. Terrell 2008 Environmental setting of human migrations in the
circum-Pacific region. Journal of Biogeography 35:1-21.
31Number 74, June 2012
Forum
Rowley-Conwy, P. 2001 Time, change and the archaeology of hunter-gatherers:
How original is the ‘original affluent society’? In C. Panter-Brick, R.H. Layton
and P. Rowley-Conwy (eds), Hunter-Gatherers: An Interdisciplinary Perspective,
pp.39-72. Biosocial Society Symposium Series 13. Cambridge: Cambridge
University Press.
Rule, S., B. Brook, S. Haberle, C.S.M. Turney, A.P. Kershaw and C.N. Johnson
2012 The aftermath of megafaunal extinction: Ecosystem transformation in
Pleistocene Australia. Science 335:1483-1486.
Summerhayes, G., M. Leavesley, A. Fairbairn, H. Mandui, J. Field, A. Ford and R.
Fullagar 2010 Human adaptation and plant use in highland New Guinea
49,000 to 44,000 years ago. Science 330:78-81.
Surovell, T.A. 2009 Toward a Behavioral Ecology of Lithic Technology. Tucson: The
University of Arizona Press.
Torrence, R. 1989 Tools as optimal solutions. In R. Torrence (ed.), Time, Energy and
Stone Tools, pp.1-6. Cambridge: Cambridge University Press.
Torrence, R. 1989 Re-tooling: Towards a behavioral theory of stone tools. In R.
Torrence (ed.), Time, Energy and Stone Tools, pp.57-66. Cambridge: Cambridge
University Press.
Torrence, R. 2001 Hunter-gatherer technology viewed at different scales. In C.
Panter-Brick, R. Layton and P. Rowley-Conwy (eds), Hunter-Gatherers: An
Interdisciplinary Perspective, pp.73-98. Cambridge: Cambridge University Press.
Veth, P., N. Stern, J. McDonald, J. Balme and I. Davidson 2011 The role of
information exchange in the colonisation of Sahul. In R. Whallon, W.A. Lovis
and R.K. Hitchcock (eds), Information and its Role in Hunter-Gatherer Bands,
pp.203-220. Los Angeles: Cotsen Institute of Archaeology Press.
White, J.P. 1977 Crude, colourless and unenterprising: Prehistorians and their
views on the stone age of Sunda and Sahul. In J. Allen, J. Golson and R. Jones
(eds), Sunda and Sahul: Prehistoric Studies in Southeast Asia, Melanesia and
Australia, pp.13-30. London: Academic Press.
White, J.P. 2011 Backed artefacts: Useful socially and operationally. Australian
Archaeology 72:67-69.
White, J.P. and J.F. O’Connell 1982 A Prehistory of Australia, New Guinea and Sahul.
Sydney: Academic Press.
Winterhalder, B., D.J. Kennett, M.N. Grote and J. Bartruff 2010 Ideal free settlement
of California’s Northern Channel Islands. Journal of Anthropological
Archaeology 29:469-490.